采样方法论文趋势 2023–2026

主要结论

• 采样方法论文从 2023 年的 122 篇增长到 2026 年的 195 篇，三年间扩大到约 1.6 倍。
• NeurIPS 是发采样方法论文最多的会议，2023–2026 年共 221 篇，占四个会议总量的 35%。
• 交叉方向里涨得最快的是MCTS：同时属于两个主题的论文从 2023 年的 4 篇增至 2026 年的 31 篇。
• 在「概率与贝叶斯」的 5 个主题里，它的 2026 年论文数排第 2；同组增长最快的高斯过程是 2.2 倍，采样方法是 1.6 倍。

代表论文

按年份和会议均匀挑选，方便快速看到每一届的样貌

• NeurIPS 2026A Model of Diverse Sampling from Language Models
• ICML 2026BFTS: Thompson Sampling with Bayesian Additive Regression Trees
• ICLR 2026Cactus: Accelerating Auto-Regressive Decoding with Constrained Acceptance Speculative Sampling
• CVPR 2026Active Inference for Micro-Gesture Recognition EFE-Guided Temporal Sampling and Adaptive
• NeurIPS 2025Antidistillation Sampling
• ICML 2025Connecting Thompson Sampling and UCB: Towards More Efficient Trade-offs Between Privacy and Regret
• ICLR 2025Bridging the Gap between Variational Inference and Stochastic Gradient MCMC in Function Space
• CVPR 2025AutoLUT: LUT-Based Image Super-Resolution with Automatic Sampling and Adaptive Residual Learning
• NeurIPS 2024In-and-Out: Algorithmic Diffusion for Sampling Convex Bodies
• ICML 2024Closing the Gap: Achieving Global Convergence (Last Iterate) of Actor-Critic under Markovian Sampling with Neural Network Parametrization
• ICLR 2024Dirichlet-based Per-Sample Weighting by Transition Matrix for Noisy Label Learning
• CVPR 2024G-FARS: Gradient-Field-based Auto-Regressive Sampling for 3D Part Grouping
• NeurIPS 2023A Dynamical System View of Langevin-Based Non-Convex Sampling
• ICML 2023Adaptive Annealed Importance Sampling with Constant Rate Progress
• ICLR 2023Any-scale Balanced Samplers for Discrete Space
• CVPR 2023Change-Aware Sampling and Contrastive Learning for Satellite Images
• NeurIPS 2026A Semantic-Sampling Framework for Evaluating Calibration in Open-Ended Question Answering
• ICML 2026Box Thirding: Anytime Best Arm Identification under Insufficient Sampling
• ICLR 2026On Coreset for LASSO Regression Problem with Sensitivity Sampling
• CVPR 2026Designing Instance-Level Sampling Schedules via REINFORCE with James-Stein Shrinkage

全部采样方法论文

共 640 篇，第 3 / 13 页

101. ICML 2026Local MAP Sampling for Diffusion Models
102. ICML 2026Markov Chain Monte Carlo without Evaluating the Target: an Auxiliary Variable Approach
103. ICML 2026MetaDNS: Enhancing Exploration in Discrete Neural Samplers via Metadynamics
104. ICML 2026Minimizing Upper Confidence Bounds: A Data-Driven Framework for Stochastic Programming
105. ICML 2026Mirror Mean-Field Langevin Dynamics
106. ICML 2026Negative Sampling From the Ground Up: A Redesign for Recommendation
107. ICML 2026Neural Quantum States in Mixed Precision
108. ICML 2026On Regret Bounds of Thompson Sampling for Bayesian Optimization
109. ICML 2026On the Robustness of Langevin Dynamics to Score Function Error
110. ICML 2026One-shot Conditional Sampling: MMD meets Nearest Neighbors
111. ICML 2026Path-dependent Discrete Amortized Inference
112. ICML 2026PCA of Probability Measures: Sparse and Dense Sampling Regimes
113. ICML 2026Prior Diffusiveness and Regret in the Linear-Gaussian Bandit
114. ICML 2026Randomized Feasibility Methods for Constrained Optimization with Adaptive Step Sizes
115. ICML 2026Reflective Hamiltonian Monte Carlo: Mixing Analysis and Application to Sampling on Stiefel Manifold
116. ICML 2026Reverse Flow Matching: A Unified Framework for Online Reinforcement Learning with Diffusion and Flow Policies
117. ICML 2026Revisiting the Volume Hypothesis
118. ICML 2026Robust Parallel Diffusion Sampling via Dynamic Jacobian Bandwidth
119. ICML 2026Robust Stochastic Gradient Posterior Sampling with Lattice Based Discretisation
120. ICML 2026SafeSpec: Fast and Safe LLM via Dynamic Reflective Sampling
121. ICML 2026Sampling and Identity-Testing Without Approximate Tensorization of Entropy
122. ICML 2026Sampling from Your Language Model One Byte at a Time
123. ICML 2026Speculative Sampling For Faster Molecular Dynamics
124. ICML 2026Stationary MMD Points
125. ICML 2026Temporal Score Rescaling for Temperature Sampling in Diffusion and Flow Models
126. ICML 2026The Loss Is Not Enough: Sampling Conditions and Inductive Bias in Contrastive Representation Learning
127. ICML 2026Thinned Mean Field Langevin Dynamics
128. ICML 2026Tilt Matching for Scalable Sampling and Fine-Tuning
129. ICML 2026Verbalized Sampling: How to Mitigate Mode Collapse and Unlock LLM Diversity
130. ICML 2026What Does Thompson Sampling Optimize?
131. ICML 2026When Sample Selection Bias Precipitates Model Collapse
132. NeurIPS 2026A Model of Diverse Sampling from Language Models
133. NeurIPS 2026A Semantic-Sampling Framework for Evaluating Calibration in Open-Ended Question Answering
134. NeurIPS 2026Adaptive Prior Selection in Gaussian Process Bandits with Thompson Sampling
135. NeurIPS 2026ART for Diffusion Sampling: A Reinforcement Learning Approach to Timestep Scheduling
136. NeurIPS 2026AT-SKM-Net: An Accelerated Trainable Sampling Kaczmarz-Motzkin Framework for Linear Hard-Constraint Feasibility on Dynamic Graphs
137. NeurIPS 2026Beyond Langevin: Sampling Multimodal Densities using the Witten Laplacian on 1-forms
138. NeurIPS 2026Breaking the $\sqrt{d}$ Communication Barrier in Federated Sampling with Adaptive Hamiltonian Monte Carlo
139. NeurIPS 2026BridgeTwist: Twisting Schrödinger Bridges for Training-Free Conditional Sampling
140. NeurIPS 2026Characterizing Memorization in Diffusion Language Models: Generalized Extraction and Sampling Effects
141. NeurIPS 2026Colored Noise Diffusion Sampling
142. NeurIPS 2026Contour Monte Carlo: Sampling via Energy Level Sets
143. NeurIPS 2026DETS: An Interval-Censored Evidential Sampling Framework for Cross-Domain Scientific Discovery
144. NeurIPS 2026Diffusion-warm sampling of the XY model enables fast thermalization at scale
145. NeurIPS 2026Does Mixed Label Imbalance Matter to Minority Collapse in Imbalanced Learning?
146. NeurIPS 2026Efficient Learning of Truncated Boolean Product Distributions: Influence to the Rescue
147. NeurIPS 2026Estimating the expected output of wide random MLPs more efficiently than sampling
148. NeurIPS 2026Finetuning with Sampling: Make SFT Generalize, Not Forget
149. NeurIPS 2026FLASH: Efficient Visuomotor Policy via Sparse Sampling
150. NeurIPS 2026GSQ: Highly-Accurate Low-Precision Scalar Quantization for LLMs via Gumbel-Softmax Sampling

常见问题

2023–2026 年四大会议一共有多少篇采样方法论文？

共 640 篇，按年份分别是 2023 年 122 篇、2024 年 156 篇、2025 年 167 篇、2026 年 195 篇。

NeurIPS 2026 有多少篇采样方法论文？

NeurIPS 2026 有 64 篇采样方法论文，上一届是 67 篇。

ICLR 2026 有多少篇采样方法论文？

ICLR 2026 有 46 篇采样方法论文，上一届是 48 篇。

ICML 2026 有多少篇采样方法论文？

ICML 2026 有 71 篇采样方法论文，上一届是 36 篇。

CVPR 2026 有多少篇采样方法论文？

CVPR 2026 有 14 篇采样方法论文，上一届是 16 篇。

哪个会议的采样方法论文最多？

NeurIPS，2023–2026 年共 221 篇。